[MLA] Loudness example with distortion on Sound Check

Sun Oct 14 19:41:23 CEST 2018

Hi Bob K,
I think you’re hearing the horrible Universal Music watermarks. Worse flutter than any analog machine!!!

Bob L.

On Oct 14, 2018, at 10:44 AM, Thomas Lund <thomas at lund.one> wrote:

Hi Bob,

I agree, and by coincidence just sent this comment for the Consumer Tech Association’s new guidelines :-)

cheers,
Thomas

> Begin forwarded message:
> 
> From: Thomas Lund <thomas.lund at genelec.com>
> Subject: [r4wg8] CTA and Dynamics Processing
> Date: 14 October 2018 at 15.11.51 GMT+2
> To: "r4wg8 at standards.cta.tech" <r4wg8 at standards.cta.tech>, Jim Starzynski <jim.starzynski at nbcuni.com>
> 
> Hello Jim, list members,
> 
> It's great to see progress, and how smart speaker manufacturers have joinined so there is a chance of avoiding new “circles of confusion” between production, distribution and consumption, to borrow an expression from Floyd Toole. With content apparently now climbing away from a low point in 2008 [Ortner], distribution and reproduction also have to cut free of those dark times.
> 
> For situations where level-normalization by itself doesn’t ensure (speech) intelligibility, our guidelines could do with a longer paragraph on dynamics processing. More the rule than the exception still, sound quality is awful across platforms, and often far from the quality of the source. Furthermore, appalling sound is getting prevalent, regardless of a decent normalization level, which for instance is the case with my favourite internet radio station, BBC Radio 2.
> 
> A new breed of ultra-harmful processing seems to be taking hold everywhere: On digital radio, internet radio, streaming and installations (malls, hotels, conference rooms etc.). Try and notice for yourself. It's like wow&flutter on steroids, a stuttering sound (albeit probably in magnitude rather than time) even more detrimental than the space monkeys of lossy codecs. Maybe it's down to a new generation of MATLAB engineers without ears, but it’s systematically destroying audio - and that deserves a clear note in a document on consumer sound technology. With Loudness and True-peak level under control, time has therefore come to focus on audio’s no 1 enemy: Junk dynamics processing.
> 
> Level-normalization based on gain reduction is the way to go, but leave further processing to experts, and design end-to-end systems for a maximum of one stage of dynamics processing between the source and the listener.
> 
> best regards,
> Thomas Lund

> On 14 Oct 2018, at 16.31, Bob Katz <bobkatz at digido.com> wrote:
> 
> On 10/13/18 4:22 PM, Ian Shepherd wrote:
>> Another example of why we need lower normalisation reference levels sooner rather than later, though...!
> 
> How about saying: "Lower normalization reference levels, no positive gain and no peak limiting." 
> 
> It's amazing how the streaming services reinvent the wheel wrong again and again and don't consult with experts like us. Even Dolby got it right historically with Dialnorm always being a downward normalization approach. Thanks to efforts of Thomas, Eelco and others, some of the services are just beginning to listen to us. 
> 
> In our recent conversations, I pointed out that it's totally impractical to expect or even talk about any streaming player applying attenuation to avoid overs just post the decode stage. Therefore even though the clipping is not baked into a well-made encoder, it has to be the responsibility of the mastering engineer to prevent decoder clipping by reducing the level that will be applied to the encoder if necessary. Mastering engineers have to bake that level into the fixed point file that we supply to the distributors. 
> 
> 
> Bob K
> 
> 
> 
> 
> 
>> Ian
>> 
>> 
>> 
>>>> On 13 Oct 2018, at 17:35, Thomas Lund <thomas at lund.one> wrote:
>>>> 
>>>> Hi Bob,
>>>> 
>>>> That’s still an excellent example, and points to the danger of allowing positive gain, especially when having a bad limiter in the system.
>>>> 
>>>> cheers,
>>>> Thomas
>>>> 
>>>> 
>>>> 
>>>> 
>>>> On 13 Oct 2018, at 18.21, Robert Ludwig <gatewaybob at mac.com> wrote:
>>>> 
>>>> Hi Ian,
>>>> 
>>>> As you like to write about Loudness I wanted to give you something I came across before you joined the group.
>>>> 
>>>> We were once on the lookout for real world examples of tracks that needed boosting on iTunes Sound Check and that boosting would cause distortion I came across this example in my iTunes library.
>>>> 
>>>> Attached is a 5:26 Victoria Mullova & Kristian Beuldenhout Beethoven Violin Sonata movement on iTunes that peaks at -2.8 dBFS on the left channel at 4:26 on the 3 loud chords.  
>>>> 
>>>> Sound Check says it needs to be raised to +8.8dB!  It technically has 6dB of distortion at that peak if Sound Check was being used.
>>>> 
>>>> BS-1770 says it has a Program Loudness of -27.7, so if iTunes had used ITU as a measurement it would have raised it even more!
>>>> 
>>>> Best,
>>>> 
>>>> Bob Ludwig
>>>> 
>>>> 
>>>> <02 Victoria Mullova & Kristian Bezuidenhout.m4a>_______________________________________________
>>>> MLA mailing list
>>>> MLA at grimmaudio.nl
>>>> http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl
>>> _______________________________________________
>>> MLA mailing list
>>> MLA at grimmaudio.nl
>>> http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl
>> _______________________________________________
>> MLA mailing list
>> MLA at grimmaudio.nl
>> http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl
> 
> -- 
> 
> If you want good sound on your album, come to
> Bob Katz 407-831-0233 DIGITAL DOMAIN MASTERING STUDIO
> Author: Mastering Audio              
> Digital Domain Website
> 
> No trees were killed in the sending of this message. However a large number
> of electrons were terribly inconvenienced.
> _______________________________________________
> MLA mailing list
> MLA at grimmaudio.nl
> http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl

_______________________________________________
MLA mailing list
MLA at grimmaudio.nl
http://grimmaudio.nl/mailman/listinfo/mla_grimmaudio.nl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://grimmaudio.nl/pipermail/mla_grimmaudio.nl/attachments/20181014/53087a9e/attachment.html>